Spatiotemporal multiscale molecular cavity visualization and visual analysis

Abstract

The analysis of molecular cavities, which are transport pathways in protein structures, is critical to the understanding of molecular phenomena. However, this work is challenging due to the high complexity and diversity of the macromolecular shapes in dynamic processes. In this paper, we propose a novel multiscale visualization method for visualizing the interaction of protein cavities. We design a series of scales and visualizations of cavities based on both temporal and spatial perspectives to allow domain experts to process their work at any scale of semantic abstraction. These scales demonstrate the chemical and structural properties of cavities and span from a complete protein to a cavity at a specific moment in temporal and spatial dimensions. We also create a continuous interaction space for multiscale applications. Finally, the applicability of our approach is proven through experimental use cases, with cavities in proteins being visualized and analyzed in a focus-and-context manner. Our collaborating domain experts confirmed that our approach is an efficient and reliable method of analyzing cavities with great potential for large dynamic cavity data analysis.